Prognosis of patients who develop cervical lymph node recurrence following curative resection for thoracic esophageal cancer

Introduction

Thoracic esophageal cancer frequently metastasizes to lymph nodes at cervical, mediastinal and abdominal regions.^1–3 Even after curative resection by esophagectomy with extended three-field lymph node dissection, approximately 50% of patients die of recurrent disease.^4–6 Recurrence in distant organs such as liver, lung and bone are generally considered to diminish the chance of extended survival or complete cure because no effective treatment for such systemic diseases is currently available. On the other hand, with loco-regional recurrence, there is a chance for complete cure with effective treatment modalities such as surgery or chemoradiotherapy, which can completely eradicate cancer cells at local recurrence sites.

Cervical lymph node recurrence is one of the common recurrence patterns in esophageal cancer and is generally considered to be distant metastasis. However, as compared with recurrence in mediastinal organs or the abdominal cavity, surgical resection for cervical node recurrence seems feasible and can be the choice of treatment. Therefore, the question which arises is whether cervical node recurrence can be cured by appropriate loco-regional treatment, if the recurrence is restricted within the cervical boundary. There have been few reports describing the prognosis and treatment of recurrent cervical nodes.

This retrospective study analyzed the clinico-pathological data of 35 patients who underwent esophagectomy and later displayed recurrence solely at cervical lymph nodes. The purpose of this study was to examine whether an aggressive treatment can give a chance for cure for patients with cervical node recurrence, and if so, what is the prognostic factor determining their survival after recurrence.

Materials and Methods

Patient eligibility

We examined the records of all 649 patients with thoracic esophageal cancer who underwent esophagectomy with lymph node dissection between January 1985 and June 2003 at the Osaka Medical Center for Cancer and Cardiovascular Diseases. Of the 649 patients, we identified 35 patients who showed solitary relapse at cervical nodes between April 1988 and September 2003. Patients having multiple site recurrence such as cervical node recurrence plus other sites of recurrence were excluded from this study. Those with the cervical node extending into the mediastinal structure were also excluded from the analysis. This study protocol was approved by the Human Ethics Review Committee of Osaka Medical Center for Cancer and Cardiovascular Diseases and a signed consent form was obtained from each patients.

Surgical procedures

Three-field lymphadenectomy was performed according to the procedures described by Akiyama et al.⁷ For lymphadenectomy of the bilateral recurrent nerve chain nodes and mediastinal nodes, thoracotomy was performed at the right fifth intercostal space through a posterolateral incision. For cervical node dissection, modified extended neck dissection was performed bilaterally. The abdominal lymph nodes that were dissected in this procedure included the right and left paracardiac nodes, perigastric lesser curvature nodes, left gastric artery nodes, common hepatic artery nodes, splenic artery nodes and celiac axis nodes. The definitions of our two-field dissection and three-field dissection procedures were as follows. The quality and extent of lymphadenectomy (including upper mediastinal, middle and lower mediastinal, and abdominal lymphadenectomy) were the same in both procedures. Regarding the upper mediastinum, both the right and left recurrent nerve chain nodes were dissected to the level of the lower edge of the thyroid gland through a thoracic approach. The three-field dissection included cervical lymphadenectomy of the paraesophageal, deep cervical and supraclavicular lymph nodes in addition to two-field dissection.

Data analysis

Hospital records were reviewed for age, sex, and clinico-pathological data regarding primary tumors such as tumor location and disease stages, details of the initial surgery including cervical dissection, cervical node metastasis, curability and preoperative chemotherapy or chemoradiotherapy. All medical charts were reviewed for details regarding recurrence nodes such as the site of recurrence, number of nodes, size and invasion to surrounding organs. Details of the treatment against the recurrence were also taken from the medical records.

Statistical methods

The survival time was calculated by the Kaplan-Meier method from the date of diagnosis of relapse at cervical nodes until the occurrence of the event or to the date of the most recent follow-up visit. Univariate analysis was done using the log-rank test and multivariate analysis was performed using the Cox proportional hazards model. P < 0.05 was considered statistically significant.

Results

Patient characteristics at prior esophagectomy

The clinico-pathological characteristics of the 35 patients at the time of prior surgery are summarized in Table 1. The mean age was 61.1 years. Tumor location was the upper third thoracic in 10, the middle third thoracic in 21 and the lower third thoracic in four. The disease stages were stage I in one, stage IIA in two, stage IIB in four, stage III in 11, stage IVA in eight and stage IVB in nine. There were only three pathologically node-negative patients (8.6%) and 15 with more than three positive nodes (42.9%). Cervical lymph node dissection was performed for 10 patients (28.6%). Preoperative therapy (chemotherapy in eight and chemoradiotherapy in three) had been given to 11 patients.

Clinical features of recurrent nodes and the initial treatment for recurrence

Table 2 summarizes the clinical features of the recurrent nodes in the 35 patients. With regard to the site of recurrence, the supraclavicular node was predominant, with solitary node recurrence more often observed than multiple recurrence (62.9%vs. 37.1%) based on physical examination and CT scan. The mean size of the recurrent node was 24.8 mm in diameter. Nodes fixed to adjacent organs accounted for 34.3%. The median interval from the date of prior surgery to the date of recurrence was 9 months. Table 2 also presents a summary of the initial treatment against cervical node recurrence. Fifteen patients were treated by radiotherapy, with a radiation dose of 60 Gy or more, except for one patient who was given 50 Gy. The clinical responses of the relapsed lymph nodes in these 15 patients were a complete response (CR) in 12, no change (NC) in two, and progressive disease (PD) in one. Of the 12 CR patients, two showed persistent CR and are still alive, and one patient, who showed local recurrence later and then underwent surgical resection, is still alive without recurrence. However, the remaining nine CR patients, two NC and one PD died due to recurrence. Of eight patients who were initially treated by chemoradiotherapy, four showed CR, one PR and three NC. Only one patient with a CR response is alive. One patient was treated with chemotherapy alone and the response was NC. Eleven patients were treated with surgical resection for recurrent diseases including two with neoadjuvant chemotherapy followed by surgery. Three patients are still alive without further adjuvant therapy. The cause of death after the above treatments against cervical node recurrence are shown in Table 3. Irrespective of the treatment modality chosen, the incidence of distant and/or lymphatic recurrence was substantially high.

Factors affecting the prognosis of patients who had relapse at cervical nodes

Figure 1 shows the overall survival of the 35 patients from the date of diagnosis of recurrence. Median survival time was 12 months with 1-year, 2-year, 3-year and 5-year survival rates of 47.2%, 26.5%, 17.7% and 8.8%, respectively. In order to examine which factor(s) determine the prognosis of patients who displayed relapse at the cervical nodes, univariate analysis was performed for clinico-pathological factors including those related to prior surgery, recurrent nodes and the initial treatment modality against recurrence. Among the factors analyzed, cervical node dissection at prior surgery (yes/no, P = 0.0178), time to recurrence (< 9 months/> 9 months, P = 0.0497) and the number of recurred nodes (solitary/multiple, P = 0.0029) were found to be significant prognostic factors (Table 4). Other factors such as age, sex, tumor location, neoadjuvant therapy, cervical metastasis at prior surgery, site of recurrence, size of recurrent nodes and whether the initial treatment for recurrence was surgery or radiotherapy, did not affect the prognosis. Next, to identify independent prognostic factors, the above three factors detected by univariate analysis were subjected to multivariate analysis. As shown in Table 5, the number of recurrent nodes (solitary/multiple) was found to be the only significant prognostic factor with an odds ratio of 2.409 and 95% confidence interval of 1.033–5.619, whereas the other two factors were not statistically significant. Survival curves were significantly different between the patients having a solitary node and those having multiple nodes (Fig. 2).

Discussion

Even after curative esophagectomy with extended lymphadenectomy, approximately 50% of patients show recurrence. Several papers have reported the recurrence rate at cervical nodes after esophagectomy.^4–6 Law et al.⁸ reported that 11% of the patients examined developed cervical lymph node recurrence after curative resection for squamous cell carcinoma of the thoracic esophagus. Katayama et al.⁹ described autopsy findings for patients who underwent curative esophagectomy for esophageal carcinoma with a metastatic rate to the cervical nodes of 11.6%. These are relatively high incidences but there have been few studies that have dealt with cervical node recurrence as a target for active treatment. Thus, the optimal treatment strategy for cervical node recurrence has not yet been established.

Our present retrospective study was conducted to examine the prognosis of 35 patients who displayed relapse at cervical nodes alone after curative surgery for esophageal cancer, and to analyze the prognostic factors that may affect patient survival. The prognosis of the 35 patients was dismal, with an overall survival of only 8.8% at 5 years and a median survival time of 12 months. However, by analyzing factors regarding prior esophagectomy, relapsed nodes and treatment against recurrence, whether the relapsed node(s) was solitary or not was found to be the only significant prognostic factor. It is generally considered that recurrence is a phenomenon in which macroscopically unidentified small cancer nodules, which had already existed at the time of surgery, grow to a visible size. Thus, recurrence at multiple nodes may suggest the existence of micrometastasis at multiple nodes, whereas solitary node recurrence may reflect the existence of micrometastasis at a single node. Considering the fact that the prognosis of patients with esophageal cancer strongly depends on the number of metastatic nodes, the results of this study are convincing.^4,5,10,11 Tachimori et al.¹² reported the survival data of 63 patients who had clinically positive cervical nodes at diagnosis and underwent surgery for thoracic esophageal cancer as an initial treatment. The 5-year survival in their study was 26.7%, which seems better than our results of 8.8% at 5 years. However, as the survival in our study was calculated not from the date of initial surgery but from the date of recurrence, these results can be considered to be comparable. For our 35 patients, 5-year survival from the initial surgery was 16.7% (data not shown).

With regard to the choice of the initial treatment modality, the prognosis did not differ between the patients treated with surgery and those treated with radiation or chemoradiotherapy, although this was a retrospective, non-randomized comparison. Actually, some of the patients who had been treated with surgery as an initial treatment against recurrence were also treated later with radiation or chemoradiotherapy when the disease relapsed again, and vice versa. Because this was a retrospective analysis of a small number of patients, a clear conclusion cannot be drawn. Thus, a prospective study with an uniform strategy such as chemoradiotherapy followed by surgery against cervical node recurrence is necessary.

Analysis of the causes of death for 28 patients after cervical recurrence revealed that only nine died of local relapse and the remaining 19 patients of distant organ recurrence, lymphatic recurrence at another site or multiple site recurrence. These results suggest that even patients who showed relapse at the cervical nodes alone may already have had disseminated micrometastases outside the cervical boundary. Such patients may need not only loco-regional therapy such as surgery or radiotherapy but also systemic therapy in order to improve their prognosis.

In conclusion, although the prognosis of patients with cervical node recurrence is poor, some of the patients still have the chance of cure. An aggressive salvage strategy should be offered for such patients. Whether the relapsed node(s) is solitary or not is the only prognostic factor and is useful in determining who is a candidate for an aggressive approach. Patients without hope for cure should be given palliative care.

References